Robot motion Under His Control
In the morning, Dr. Yonglin Chi, the CEO of GENE automation, took the time to accept interviews from several medias including Science and Technology Daily, Robot Lecture Hall, Ofweek Robot Network, Robot Library, and China Robot Summit. And explained robot control technology -GENE robot controller theory in a simple way.
Profile of Dr. Yonglin Chi
Dr. Yonglin Chi, Chief Scientist and CEO of GENE Group, has been engaged in industrial robot motion control for many years. He has witnessed the development of China's robot industry for many years. With his keen business vision and rigorous scientific research attitude, he has become one of the main founders of GENE Group. The international robot giant Kawasaki robot used the "Chinese brain" – GENE controller -a smart robot motion controller that compares the performance of the international robot controller.
Dr. Yonglin Chi, has been in the robot industry for more than ten years. He experienced the rapid development of the world's top robotics companies in China, participated in controller-related product and project research and development, and has also done a lot in robot application promotion.
Controller advantage: rooted in the local
For the research and development results of GENE team, Dr. Chi said: "We have inherited the advantages of advanced foreign controllers nowadays, and based on the demand of China domestic robots, developed GENE robot controller with completely independent intellectual property rights. On this basis, Completed the applications of six-axis robots, seven-axis single-arm robots, and seven-axis dual-arm robots."
GENE robot controller is closely integrated with China local needs, which results the deep cooperation with Kawasaki. By using GENE controller to lead Kawasaki robotic arms, a series of unique robot products has been formed. Why the international robot giant uses the "Chinese brain"? And what are the advantages of this robot product series? Dr. Chi answered our confusion in great detail.
The new product series is based on accumulated experience of Kawasaki robot body and quality and intelligent features of GENE controller. It’s a high-performing robot welding and cutting system at technical level.
The GENE team is a down-to-earth team in contact with real demands of industry frontier. Its technology in fields of intelligence and sensing has developed rapidly. The robot controller grasps the trend of development and combined to actual requirements, and was adapted to new and special applications. Comparing to foreign robot brand, the intelligence of GENE controller become the most competitive advantage, especially for special application development, in the local market, it surpasses the first-class international brands.
GENE robot systems also have a strong service advantage. Foreign companies only sell robots in China, and R&D is far away from the frontier. GENE R&D and service personnel are completely integrated, and can truly serve customers based on their needs."
Dr. Chi said with great confidence: "Our controller and Kawasaki robotic arm combined product series is a very strong product series that will be intensively cultivated and rooted in China market."
Robot controller features: deep embedded intelligent manufacturing
Nowadays, China is facing a great development opportunity -the universally advocated intelligent manufacturing (IM). However, IM is not simple, and is far from being solved by adding a little bit of interconnection and adding a little intelligence to the existing manufacturing base.
The true meaning of intelligent manufacturing (IM) needs to go deep into the first-line of manufacturing industry, and dig deeper into the difficulties and key issues to be overcome in modern manufacturing. The traditional manufacturing industry is with a relatively small variety but large-scales item production, like automobiles with production capacity of similar models exceeding 300,000 units per year. However, in current manufacturing industry, there are many varieties while with small quantity. This is the development trend at present. Under the premise of this development demand, the existing manufacturing model must make some changes.
The most prominent feature of the robot application series developed by GENE Group is the deep embedding of intelligent manufacturing features at the robot controller level.
In the introduction of Dr. Chi, we learned how the controller can make the production line with key features of adapting to many varieties and small batches in the intelligent manufacturing process.
The first feature is the motion intelligence that adapts to product diversity.
This raises a request on the controller -good sensor-based identification and compensation capabilities. People can easily identify a variety of products, and they can understand it with their eyes, but today's production lines or manufacturing equipment are far from this level.
Therefore, GENE embeds the function of depth sensing fusion in the controller, which can quickly connect to various sensors such as vision, force, laser, acceleration, etc., and identify product differences on the production line to adapt to the difference of incoming materials.
The second feature is intelligent processing. Even if the product type and quality change, the robot system can automatically identify and compensate for these changes to ensure the performance and consistency of final products.
Therefore, GENE team embedded big data in the controller, which can automatically optimize and improve the entire manufacturing process according to the historical big data in the manufacturing process, the data information and sensing information of the current incoming product. Material changes, or some subtle differences in incoming items, can also be automatically identified. This is the real intelligent processing.
As we all know, the manufacturing competition is very fierce now, and each production line requires extremely high production efficiency. Under the requirements of multi-variety and small-volume manufacturing environment, the production scheduling and decision-making of the production line can be said to be the key factor determining the efficiency of the production line.
GENE controller has built-in intelligence module for distributed decision-making, which changes the traditional manufacturing industry with a hierarchical structure of ERP, MES to workstation. The robot controller collects and stores massive information of current work, and can also access to all relative data of production line through distributed data. Based on the global relevant information, it can make the corresponding key decisions at the workstation or robot controller level.
The biggest difficulty faced in production is the rapid change of product demand, which leads to short product life cycle and rapid switching of production orders. When the product changes, the decision must be made from ERP, MES to the whole execution system project -a complicate operation. So many factories relays on artificial operation, or a specific complicate system.
With a distributed decision-making system, the biggest change is: where things happen, where decisions are made, from the bottom up to confirm the key issues of the switch to make decisions, rather than the top-down decision-making model. This is the intelligence of the very important production decisions GENE embed in the controller.
In fact, the core of intelligent manufacturing (IM) is to bring value to users. Users not only need a standard product, but also some special customized services. GENE team has embedded big data acquisition and recording functions to each robot controller and robot application systems.
After the production’s completion, the key manufacturing data and service data of each product can be concentrated in service provider; the customized service and product information from end user could be collected through Internet. In this way, the base of promotion for smart products with customized services are laid.
It can be said that GENE embeds a variety of new methods and modes for intelligent manufacturing in the controller and series of robots, which plays a positive role in promoting the rapid development of intelligent manufacturing and intelligent products in China.
"For the core components made in China, we provide a lot of interface and compensation technology in our robot control system. The overall performance and reliability of robots are improved. Even with domestic components in robot controller, it achieves global levels with performance comparable to foreign components.”
"In this way, the entire Chinese robot industry has the ability to compete with the world's major robot families. This is a goal we make. GENE cooperate with companies from Yangtze River Delta and other domestic companies at the robot controller to influence and improve the whole robot industry. In China, we must give full play to our industrial advantages to reach the level of competition with the four major families. This is the status and real power of China entire manufacturing industry. By doing this, China can truly become a manufacturing country."
Controller R&D ideas: originating from problems, and to solving problems
A reporter asked: "Sensor fusion, embedded big data, distributed decision-making, are the three groundbreaking researches of controllers. Are they based on existing technology integration or an innovative model? Since the three have achieved great results, what kind of thinking we have on R&D base?
The answers from Dr. Chi is:
Sensor fusion is adapting to the development trend. Nowadays, robot research wants to integrate more sensor information into various control and decision-making of robots. In this respect, what is the biggest breakthrough and technical advantage of GENE?
The general convergence is based on interaction or in the case where the application layer of the control system is not real-time. Dr. Chi believes that the breakthrough in sensor fusion in the future is in real-time response. Why we regard human as the most intelligent? For example, when fire burns or pin needles, human brain will react immediately. This is real-time reaction in control field.
With sensor information, the reaction will be done in milliseconds. And the next step will be in microseconds for us. This is a big breakthrough we made in sensor fusion. In industrial competition, the performance and functionality competition come down to time response eventually.”
A very simple example, there are two control methods for robot to grab a bottle of water. One is that the sensor has a separate sensing processor that transfers the visual information or processed position information to the robot. The robot controller determines robot react position and perform motion based on the information received. It is the interaction of two heads-the sensor processor and robot controller.
This kind of interaction is usually between second and hundred milliseconds. It is definitely not as natural and smooth as human beings. If the robot reacts like a human being, it must get through response time progress or motion control treatment limit and lay intelligent sensor fusion technology at the underlying of motion control and planning. GENE controller breakthrough is based on the existing foundations of the industry and own understanding of the underlying motion control. Dr. Chi Yonglin said.
The other is a breakthrough in processing technology, rooted in the understanding of the entire industry in GENE team. On the production line, especially in many assembly lines and manufacturing lines in China, the most typical outstanding feature is that the incoming materials are different. Dr. Chi cited an example of a household circuit breaker.
During the assembly process of circuit breaker, there are lots of small parts. In China, four or five operators sit on a table to complete the assembly of dozens of parts in less than 30 seconds. In the process, every difference in incoming materials will be identify swiftly before assemble. However, the current level of robot is far from this intelligence. Our team has been exposed to a lot of similar needs, and these requirements have led to a very core question: how the machine adapts to these differences.
Much of the intelligence of these things is not a purely motion issue. It requires prior information and historical experience information for all processes and unknown parts. These problems can only be solved by combining the information and movement process in deep. Therefore, the big data and processing technology were put into the controller, in order to solve the situation that the parts are different and the process needs to be continuously optimized and improved. This is something truly creative and stems from the problem.
The third distributed decision.
Now the bigger and more complex the manufacturing system is, the harder the decision is to make and at higher manufacturing cost. If a workstation at the bottom of the production line was changed, there will be a big change in the entire production process. The traditional way requires all the information and relative data of this workstation to quickly released to the top of the manufacturing system in order to make the right decision. This decision-making model determines that the flexibility and adaptability of the production line will not meet some of the current needs.
What should we do with such a problem?
In labor gathering factories, once operator replacement was done in a certain process, is it necessary to report the changes? There may be some different, report is not necessary. The newly changed person can distinguish the differences of coming parts, the direct leader knows that he is doing the job. Making the decision locally will be enough.
Imitating people-oriented manufacturing decisions, GENE team only realize where the problem occurred and where to make decisions, and established such a distributed decision-making framework for manufacturing systems. The underlying robot controller of the production line can get all the relevant information needed for its work, know the current work and the next step of work, self-decision, and report the decision result to the upper layer, so that the upper system can do the overall coordination.
In this way, lower layer feeds back the results of the underlying distributed decision to the upper layer, and the upper layer makes decisions based on the underlying information. This decision-making approach is more adaptable to modern manufacturing flexibility. Therefore, the decision of GENE controller is to imitate the way people organize the society and make a change in decision-making, called "distributed decision-making".
2017-09-22 Reprinted from China Robot Summit (click to check Chinese Version)